Connecting to Learn

Is a network capable of learning? Can learning reside in a network? According to connectivism, the answer to both of these questions is “yes.” Check out George Siemens’ views in this video. These arguments seem to me to be counterintuitive. How do other more established learning theories approach this issue?

Constructivism would say that the learning is a result of learner exploration and discovery, of trial and error, of guided social interaction within a learning environment. However, social constructivism (a flavor of constructivism) would find learning in the midst of social networks. The question is “If a network is not social, can learning reside there?” Or “If a network is not presently being used by an individual or group can learning be said to reside on the network still?”

Perhaps cognitivism is the learning theory closest to connectivism’s views on learning and networks (though I certainly doubt cognitivists would agree with my assessment). I say this because cognitivists often view learning as memory acquisition, sythesis, and development with an output of changed skills, knowledge, and attitudes. In this view, neural networks are one of the most important bases of learning. How do neural networks learn? One might answer “the same way any other network can be said to learn.” What is the difference between data being stored and manipulated in a brain and data being manipulated and stored in a machine or on a jump drive? Can my jump drive learn then?  I think data and information are being confused here with knowledge and learning. Technology may help humans communicate, store, reorder, represent, and organize information or even sometimes knowledge (take Wikipedia for instance), but is the technological network the tool, is it the environment, or is it the basis of learning (or is it even learning itself)?

Behaviorists will only talk about what a person can see in a person’s behavior as the way to evaluate learning. But on this basis, can not a computer learn as well? It may give the same output as a human. For instance, videotape an instructor, host the video on YouTube, then play the video on your computer. According to behaviorism, hasn’t the computer and the network just learned? After all, it showed that it learned the instructor’s content perfectly by spewing out a carbon copy. It mimiced the behavior that the instructor was seeking to assess. Voila, learning.

Perhaps what connectivism does best (at least in my very narrow opinion) is that it points out a very common weakness in modern theories of learning: philosophical materialism. If what cannot be measured cannot be said to exist (positivism), and if humans are no more than phyical (philosophical materialism), and if learning is therefore merely a physical phenomenon not embued with any true and eternal significance (it is merely changed behavior, memory, or society), can not learning be said to reside also in our less organic counterparts–the computers that inhabit our homes and offices and the networks that tie them and us together? Isn’t this the basis of the idea of “artificial intelligence?”